SIsignalshq
DOCUMENT

Retry extraction for a failed/stuck document

Re-run extraction for a document whose extraction FAILED or got stuck (no review workflow / no editable fields). Creates a NEW extraction job for the document and re-extracts with its canonical type (or re-classifies if it was never classified). Idempotent under concurrency: a genuinely in-flight job is not duplicated, and a job left non-terminal by a killed worker stops blocking once it ages out. Does NOT change the document's type or supersede a prior review.

Read the response body, not just the status. This returns 202 whether or not work was enqueued; status: SKIPPED with the pre-existing extraction_job_id means the guard considered a job already in flight and nothing was queued.

force=true bypasses that guard outright — for a document an operator knows is dead but which has not yet aged out. Safe under concurrency: the extract task's execution claim makes a second dispatch skip rather than double-extract.

post/internal/api/document/{document_id}/retry

Request

  • Base URL
    https://api.signalshq.io (the document declares no server; this is the origin it was published from)
  • Auth
    OAuth 2
    Password
    Token URL: /auth/login

Path parameters

document_idstring uuid required

Document ID to retry extraction for

Query parameters

forceboolean

Bypass the in-flight guard and always create a new job. Use when a document is stuck behind a job whose worker died and the normal retry returns SKIPPED.

Request body

client_idstring uuid required

Client ID for the document

Response

Successful Response

document_idstring uuid required

Document ID

extraction_job_idstring uuid required

Extraction job ID for polling status

statusstring required

Initial job status (e.g. ACCEPTED)

Changes

No recorded changes to this operation across all 1 revision of this API.